Free and Bound Water in Elastic and Non-Elastic Gels - The Journal of

F R E E AND BOUND WATER I N ELASTIC AND NON-ELASTIC GELS* BY I V A N D. J O N E S AXD ROSS AIKEN GORTKER

Introduction The more recent trend in all studies of colloidal systems has been toward an application of physico-chemical methods, in order that the reactions which take place under given conditions may be defined more exactly and be determined quantitatively. Later investigations in the narrow field of studies on the binding of water by substances in the colloidal state have been in keeping with this modern approach to problems in colloid chemistry. The ability of certain substances in the colloidal state, when in the presence of a given liquid, to hold a quantity of that liquid with great forces of attraction has been long recognized. The exact nature and intensity of these attractive forces have, to date, not been clearly defined. This attraction, when exhibited toward water, results actually in a hydration of the colloidal particles and gives rise to the phenomenon known as water-binding. Water-binding or hydration is accompanied with certain physical and chemical changes in the colloidal 3ystem. Thus, the viscosity is relatively greatly increased, the vapor pressure may be markedly reduced, and a certain portion of the water in the system will no longer act as a solvent. Foote and S a ~ t o n ~ concluded ~ J ~ J ~ that a definite portion of water would not freeze from the inorganic hydrogels which they studied and the investigations of Rubners2 and R o b i n s ~ were n ~ ~based ~ ~ ~on~ the ~ ~hypothesis that a t temperatures as low as --20°C bound water would not freeze. I t is the purpose of the present paper to report a study of the relation of temperature to the quantities of water which appear to exist in the “free” condition, and in the “bound” or “unfree” condition a t known temperatures below the freezing point of a given colloidal system. Also, the behavior upon repeated freezings a t low temperatures of certain organic hydrosols and hydrogels is compared with that of certain inorganic colloidal systems. Historical Since the early investigations of colloidal systems were generally of a qualitative nature and only in the more recent studies has quantitative technic been introduced, in reviewing the literature dealing with the effects produced on colloidal systems by freezing, we will consider first qualitative and later quantitative observations. *Published as Journal Series No. 1023, Minnesota AgriculturalExperiment Station. Condensed from a thesis presented by Ivan D. Jones to the Graduate School of the University of Minnesota in artinl ful6llment of the requirements of the Degree of Doctor of Philosophy, March 1931. &e authors wish to express their sincere ap reciation to Dr.J. H. Beaumont, head of the Department of Horticulture of the North 8arolina Agricultural Experiment Station and to the authorities of the North Carolina Agricultural Experiment Station for permission to utilize laboratories which enabled one of ua (J.) to complete certain of the experiments reported in this study.

388

IVAN D. JONES AND ROSS AIKEN GORTNER

Qualitative studies o n the effect oj freezing o n colloidal systems.-Records are available of observations on the effect of freezing of colloidal substances which antedate the recognition of the phenomena of the colloidal state by Thomas Graham. Thus in 1820 VogeP3noted that a distinct change in wheat starch paste occurred upon freezing. When the frozen paste was thawed, a liquid containing but little dissolved substance separated from a spongy mass. MolischS6 in 1897 reported a microscopical study of the freezing of a zy’, gelatin sol. He found that on freezing, the gelatin formed a net-work, the meshes of the net being filled with ice-masses and air bubbles. Upon thawing the frozen mass, he observed that the gelatin possessed a sponge-like appearance. Similar studies, upon gum tragacanth, gum arabic, egg albumin, and tannins were made by Molisch. Bobertag, Feist and Fischer,2 and Fischer and BobertagI2 considered that freezing produced, in hydrosols of gelatin, agar-agar, soap, and carragheen moss changes which were only partially or very slowly reversible upon thawing. Callow,’ in a study of the rate of ice crystallization through super-cooled gelatin gels, maintained that the “separation of ice irreversibly ruptures the structure of the gel.” This conclusion was drawn from observing the behavior of the gels upon repeated freezings. Inorganic hydrosols have been studied by many workers. The results obtained have not been in complete agreement. Bobertag, Feist and Fischer? observed that upon the thawing of ferric hydroxide sols, frozen a t - IOOC and -70°C, the sol was re-formed. LottermoseP observed that this sol, dialyzed until it had a specific conductivity near that of water, was completely precipitated by freezing. He concluded that the electrolyte present in an undialyzed sol served as a stabilizing agent. By the freezing of incompletely dialyzed ferric hydroxide sols, he was able to effect only a partial precipitation of the colloidal material. Lottermoser and also BrunP independently observed the effect of freezing upon the hydrosol of silicic acid, and reported that upon thawing the frozen sol it separated into a clear liquid and a precipitate. VanzettP from repeated experiments with silicic acid gel frozen a t temperatures as low as -zoo°C concluded that the final composition of the gel was independent of the temperature of freezing. A certain portion of the water present could not be removed by freezing. Liesegang30 repeated Molisch’s experiment with a 2% gelatin sol and reported very different results. He found that the water on freezing formed fernlike ice patterns. On thawing, the patterns remained in finest details, and he adds that themost gelatinwas found where the most ice had appeared. Apparently, therefore, water was not separated from the gelatin upon freezing and subsequent thawing. Stiles66 has raised the question as to whether the differences in results obtained by Molisch and by Liesegang were not caused by different rates of freezing of the gelatin sol.

FREE AND BOUND WATER IN GELS

389

I n an excellent treatment of the fundamental physical and chemical principles of the freezing process StileP has definitely shown that the conditions under which freezing takes place determine, to a very large degree, the effects which will be produced by freezing. His work was based on the early investigations of T a m m a n ~ Pwho ~ concluded that in the crystallization of a super-cooled liquid, two factors were important, (A) velocity of crystallization and (B) the number of centers from which crystallization takes place. Moran3’ in microscopical studies of the effect of freezing on gelatin gels, found that the structures produced in the gel depended upon the rate of freezing and the concentration of the gel. It is of particular interest that Moran demonstrated that centers of crystallization could be either external or internal, this condition being determined by the temperature a t which freezing began. Hardyz1 carried on further microscopical investigations with frozen gelatin gels, and states “so far as my observations go, when crystals (internal) of pure ice melt, the water is reabsorbed a t once by the surrounding gel, leaving only a tiny cleft.” These studies of Stiles, Moran, and Hardy offer an explanation of Liesegang’s observations which were a t variance with the earlier studies of Molisch. hard^'^**^ advanced the theory that with gels or sols, dehydration could be considered a reversible process, if a gel or sol resulted normally from the addition of a colloidal substance to water; and that it would be an irreversible process, if special conditions were required to produce the gel or sol. Fischer“ in 1911pointed out that the freezing process might similarly be considered as either reversible or irreversible, since freezing is only a special type of dehydration. Quantitative Studies on the E$ect of Freezing on ColloidalSystems.-MullerThurgau3*in 1880published the results of investigations on apple and potato tissue. In his studies the material was first frozen in an ice and salt bath and then introduced into a water calorimeter. From the quantity of heat required to melt the ice in the sample, he was able to calculate the quantity of water which had been frozen. As has been noted by A. Kuhn,*g MullerThurgau’s first results were undoubtedly erroneous, for he gave no consideration to the molecularly dissolved substances in his samples. The principle involved in Muller-Thurgau’s studies has, however, been of great value in that it laid the foundation for much biological research. Foote and S a x t ~ n ~introduced ~ J ~ J ~ dilatometric technic into studies of the freezing of colloids. They chose as systems for investigations, the inorganic gels of silica, alumina, colloidal ferric hydroxide, and a mixture of lampblack and water. With every system studied they found that a certain portion of the water present remained unfrozen. They accordingly concluded that water in the inorganic hydrogels existed in three forms: (A). Free water-the water which froze between the temperatures o°C and -6OC. (B). Capillary waterthat portion of the water in the sample which froze a t temperatures below -6°C. (C). Combined water-the portion of the water which represented the dBerence between the total water in the sample and the water which could be frozen.

3 90


Applications of the dilatometer method have been made by several investigators, notably by B o u y o u ~ o s in ~ - ~soil studies, McCool and Millar33 in soil and plant relationships, and by and L o t P in winter-hardiness studies. Parker,45in 1 9 2 I, reported some exceedingly important observations on the effect of finely divided insoluble material on the freezing point of different liquids. He found that with water the magnitude of the freezing point depression depends upon at least two factors, (I) the relative surface area of the solid, or the degree of subdivision of the material and (2) the affinity which exists between solid material and water. Newton and Gortnei3 in 1 9 2 2 suggested a new method for the quantitative measurement of water held in the “unfree” or bound condition in sols of hydrophilic colloids. This method is based on the hypothesis that a certain portion of the water in a hydrophilic sol is associated with the colloid in such a form that it will no longer act as a solvent, as contrasted with “free” water. The addition, then, of a definite quantity of a soluble material to a sample containing a known amount of water would cause a lowering of the freezing point of the mixture, in proportion to the molar concentration which resulted, and a positive deviation of the observed from the theoretical freezing point depression would be a measure of the bound-water which was present. Kruyt and Winkler28have verified the findings of Newton and Gortner regarding the effect of the presence of hydrophilic colloids on the freezing point depression of solutions of molecularly dissolved substances. The Newton-Gortner method has served as a quantitative measure of bound-water in many plant investigational projects and related studies, e.g. N e ~ t o n , 3 9 ~Harris ~ ~ ~ ~et1 allz2Newton and Cook,42 Newton and Martin,44 Martin136and Crist.8 Kuhn29 cites an extensive study made by Rubner of water-binding in Laminaria and materials of animal origin. R u b n e F employed a method very similar to that first used by Muller-Thurgau, and considered the latent heat of fusion of ice to be a measure of the quantity of water which could be frozen. ThoenedG0studies on the effect of freezing upon water-binding in gels and animal tissues involved an experimental procedure modeled closely after that of Rubner. Thoenes concluded that the method yielded reproducible values when dealing with animal tissues but that these values were only relative. It gave a measure of water-binding intensity but not a true measure of the quantity of water which existed in the bound form in the living organism. Robinson,48 in a study of winter-hardiness in insects, employed the method described by Thoenes. He improved the technic49 and pointed out that this method yields only minimal values for bound-water. He demonstrated in a striking manner the importance of bound-water studies in animal physiology. Moran37 made quantitative investigations on the effect produced by freezing gelatin gels a t different temperatures. Having noted that ice formation was entirely external under certain conditions of freezing, Moran

FREE AND BOCND WATER I N GELS

391

took advantage of the fact that the ice appearing in such manner could be removed from the resulting partially dehydrated gel sample, and thereby he determined the phase equilibrium between ice and gel a t different temperatures, He found the equilibrium completely reversible in the gel concentrations which he studied. Hill,?3 and Hill and KupalovZ4have recently studied the bound-water problem in animal tissues using the vapor pressure method. They arrive a t results which are not in agreement with those reported by ThoenesJG0who found relatively large amounts of bound-water to be associated with the hydrophilic colloids of muscle, whereas Hill finds very little if any bound water in either blood or muscle. I t seems probable that this finding may in part be explained by the technic used and by the assumptions which Hill made. Since Briggs5 is presenting on this same program an extensive study of the vapor-pressure method, the reader is referred to his paper for a discussion as to the possible causes of the divergent conclusions noted above. That the majority of physiologists, physicists and physical chemists agree that some sort of water binding occurs in hydrophilic colloid systems is indicated by the extended discussion which was provoked following the recent presentation before the Faraday Society of a summarizing paper on “bound” water by Gortner.l8 Experimental T ~Problem-It P is apparent from the preceding historicalreview that there is not complete agreement as to the effects which are produced when colloidal systems are frozen. Similar lack of agreement is found when dealing with methods of measurement of bound water. Freezing of colloidal systems has been employed by some investigators as a method in studying the water-binding in certain materials. Foote and Saxton,13l4 !& Rubner,j* and Robinson**4 9 , 5 0 considered the freezing method as a quantitative measure of water-binding capacity in the materials under investigation. ThoenesG0concluded from a study made by means of the Rubner method that it did not give absolute values for waterbinding capacity. K ~ h n ’has ~ emphasized, in a very complete review of water binding in colloids, that freezing reduces water-binding in colloidal systems, also, that any freezing method is reliable only to the extent that it measures the intensity with which water is bound, and, accordingly, can not be considered a quantitative measure of water-binding capacity. Gortner!’ (p. 2 2 7 et seq), in a general discussion of the problem, suggests that the bound water is probably present in the form of oriented dipoles and adds1*that, “I do not believe there is any sharp line of demarcation between “free” water and “bound” water, but that wemust postulate aninsensible gradation between molecules of water having the normal activity of pure water and molecules of water where this activity has been so reduced that such molecules have become to all intents and purposes a part of the solid upon which they are adsorbed. One method of measurement may be sensitive enough to differentiate between water molecules having a given activity and those molecules having a greater activity. Another method of measurement

392

IVAN D. JOSES AND ROSS AIKEN GORTNER

may give somewhat different results because it draws the dividing line a t different activity values. What we in biochemistry need at the present time are extensive series of measurements carried out on many biocolloid systems by many different techniques. When these data have all been accumulated, then perhaps it will be time enough to attempt to theorize as to the exact nature of “bound” water.” The pvrpose of the present paper was to determine with certainty whether or not “bound-water” would freeze, and if so, a t what temperature this freezing would take place. If bound-water would not freeze, this process could be employed as a quantitative measure of the bound-water content of hydrophilic colloidal systems. If bound-water would freeze, it was considered that a relationship should exist between the rupture of the gel structure and the freezing of bound-water. Newton and Martin,44employing the Newton-Gortner method of boundwater measurement with different concentrations of organic hydrosols, found that with increasing concentration of the colloid, there was an increase in the quantity of water which was “bound.” This is taken as evidence to support the supposition that gel formation results when the bound-water represents an appreciable portion of the total water present. If then, the bound-water content can be reduced by freezing, a partial or complete rupture of the characteristic gel structure exhibited by certain colloidal systems should result. The Method.-In order to observe the behavior of hydrophilic colloidal systems upon freezing, it was desirable to employ a method which would permit continuous readings to be taken on a given sample at any desired temperature. Such a method should permit a study of the behavior of the system as the temperature was lowered below its freezing point and also as the temperature was raised again to the melting point of the ice. The dilatometer method was accordingly chosen. The principle of the dilatometer method is well known and may be summarized as follows: I n any given system changes in temperature are accompanied by corresponding density changes of the system. If with a temperature change, there occurs a liquid e solid transition, this is, in general, accompanied by a corresponding change in density. If the mass of the system be fixed, the density changes may be followed by the volume changes which occur. With water, solidification takes place a t 0°C with a density changez5from 0.9999 to 0.9168. This density change results in a volume increase of 0.090pq cc. per gram of water. Therefore, by measuring the expansion produced by freezing, it is possible to calculate the quantity of water which was frozen. If freezing occurs gradually as the temperature is lowered, there will occur simultaneously (A) an expansion resulting from the gradual formation of ice and (B) a contraction resulting from the temperature lowering of the entire system. Knowing the density change for all solids and liquids in the system for any given temperature change, the ice formation which results from a given temperature change can be calculated.

FREE AND BOCSD WATER I N GELS

393

As has been indicated, other workers have used dilatometric technic in studying ice formation in colloidal systems. They have, in general, calculated the amount of water in the system from a moisture determination of the sample under investigation. I n addition the quantity of water which was frozen a t a given temperature has, in general, been calculated directly from the expansion which resulted from the freezing process. I n the greater part of this investigation, a known quantity of pure water was pipetted into the dilatometer, and the volume change produced by cooling the system to definitely recorded low temperatures was measured. Following the determination of such “water curves” the same quantity of water was added to small amounts of the solid material under investigation in the dilatometer, and the expansion and contraction resulting from exposures to low temperatures were again measured. Thus, in the second series of measurements the conditions were identical with those prevailing in the first, the only difference in the system being the small quantity of material under observation. I t was therefore assumed that the two resulting curves were directly comparable, and if the material did not have the ability to “bind” water, that the second curve would coincide exactly with the first one. Further, it was considered that the measureof the quantity of the water which the material would bind would be indicated by the extent to which the two curves deviated from each other. It was expected that gradually lowering the temperature would cause a convergence of the curve resulting from the freezing of the colloidal system toward the curve for pure water. If such convergence did not occur, then one could rightly assume that decreasing The temperatures did not alter the amount of bound water. Stated as used in this study differently, if at the point of initial freezing all water which would freeze was transformed into ice, then the curves resulting from plotting the readings obtained at different temperatures below that of freezing would be a line parallel to that given by the pure water sample. Any deviation from such a parallel line would be expected to take place in the direction of the pure water line, as it would thereby indicate a decrease in the total quantity of water held in the non-frozen state. Descrzption and calibration of the apparatus.-The apparatus consisted of a bulb-capillary-stem dilatometer mounted on an engraved meter stick as shown in Fig. I , and a low temperature thermostat in which the desired low temperature could be produced and maintained. The dilatometer figured was the most satisfactory and most sensitive one of many designed for these studies. The bulb of the instrument was blown from pyrex tubing about 0.j X 3 cm. One end was sealed shut; the other end was fitted with a carefully ground glass stopper which was held firmly in place by means of rubber bands. The stem about 5 5 cm. long was of pyrex capillary tubing having a I mm. bore and was attached to the center of the bulb at a direction perpendicular to its long dimension. I

1,

:&Ameter

394


The dilatometer had a capacity of about 5 cc. The sample taken usually occupied a volume of from I to 3 cc. The remainder of volume of the dilatometer bulb and a portion of the stem were filled with pure toluene. Toluene was chosen because of its known coefficient of thermal expansion,Zj its high boiling point, and its inactivity toward aqueous colloidal systems. The high coefficient of thermal expansion of toluene produced a decided density change with a small temperature change, therefore serving to counterbalance the expansion produced by the freezing of the water in the sample and made possible the use of a relatively short stem for the dilatometer. The stem of the dilatometer was attached perpendicularly to the bulb primarily to avoid a possible fracture of the bulb resulting from the expansion of the sample when it froze, and also to facilitate the filling the dilatometer, mixing the sample, and freeing the system from air bubbles. The volume of the dilatometer was calibrated at room temperature by means of weighing the mercury it would contain. The diameter of the capillary stem and the uniformity of its bore were determined by measuring with a microscope equipped with a movable stage the length of a weighed mercury thread in consecutive sections of the tube. All variations in the bore of the selected capillary were found to be well within the limits of experimental error. The actual diameter of the uniform capillary was further checked by weighing a thread of mercury which filled the greater portion of the length of the tube. The capillary chosen had a volume of 0.0096078 cc. per cm. length. I t was possible to read the length of the toluene column in the capillary with an accuracy of 0.05 cm., thus permitting an accurate measurement of a volume change as small as 0.00048 cc. The low temperature thermostat consisted of a well-insulated doublewalled container, the inner vessel having a capacity of approximately I liter. The cooling medium was alcohol, the temperature of which was lowered as desired, by adding a sufficient amount of solid carbon dioxide, similar to the technic suggested by Dunn.8 In order to assure uniformity of temperature in the bath the alcohol was stirred continuously by a small electric motor. I t was essential to maintain a uniform temperature in the bath, since the temperature of the bath was taken to be the temperature of the system inside the dilatometer after the contents of the dilatometer had arrived at a temperature equilibrium. The size and construction of the dilatometer prohibited the use of a thermometer inside of this instrument. A 0°C to -6s°C pentane thermometer was used in determining the temperatures. This thermometer was checked against a thermometer which had been standardized from a Bureau of Standards certified instrument. Materials. The materials employed in this investigation fall into two classes. (A) those substances which form gels of the elastic type and (B) those which form inelastic gels. Gelatin was chosen as a typical substance which forms an elastic gel. Gelatin was especially suitable from many standpoints, two of the factors being its behavior in cold water and its ready sol % gel transformation.


395

Since the method of mixing water with the dry sample was that of adding a definite and constant quantity of water to the dry material in the dilatometer, it was essential that the finely divided substance should be readily wetted in a uniform manner. Gelatin behaves satisfactorily when placed in cold water but most other dry hydrophilic substances tend t o form a lump, the interior of which is only very slowly wetted. After the initial swelling of the gelatin in cold water had taken place, the dilatometer could be immersed in warm water, this causing the formation of a sol which upon cooling would set to a gel. The other substance chosen as an example of the elastic gel type was the thick portion of the white from fresh eggs. Materials of the inelastic-gel type were activated silica gel obtained from the Silica Gel Corporation, and the colloidal coagulum of ferric hydroxide.

Experimental Data The Volume Change of Toluene with Temperature.-In order to test the accuracy of the dilatometer, a study was made of the behavior of pure toluene upon exposure to temperature changes. The dilatometer was completely filled with toluene, and the readings taken at different temperatures were plotted. The resulting curve is shown in Fig. 2 . All readings fell on a straight line and the values were easily reproduced in repeated temperature-lowering or temperature-raising cycles. When this line was checked against the esisting data on the contraction of toluene2: our values proved to be about 2 . ~ too high. Yo reason was found for this discrepancy. Table I gives the readings from which the toluene curve was constructed. In recording the data taken in these studies, it was considered, in every case, that the toluene column had zero length when the system was in equilibrium at 0°C. Therefore the readings which resulted from temperaturelowering are recorded as cm. contraction or as a negative length of the toluene column. If then at any temperature an expansion occurred which caused the toluene column to rise above the point at which it stood at o°C, the readings became positive in sign. I t will be seen that this change in sign of toluene length occurred in every case where water was frozen. The Volume Change of the System, Toluene-Water, with Temperature Change.-Fig. z phows the form of the curve which results from freezing a sample of pure water. In this instance 1.955 grams of distilled water was frozen. The line AB represents the contraction of the toluene-water system between the temperatures oo C and - 11.1' C ; line BC represents the elongation of the toluene column in the capillary stem of the dilatometer due to expansion of the system produced by separation of ice at a constant temperature; line CD then represents the contraction of the system, toluene-ice. It will be observed that the curve resulting from the freezing of the water sample is well represented by a straight line, as is indicated by readings taken both during temperature-lowering and temperature-raising.

7

~

IVAN D. JONES AND ROSS AIKEN GORTNEH

396

Fig. 2 represents but one of many determinations, all of which were in exceedingly close agreement. Accordingly, the curves of Fig. 2 have been used as reference curves in later studies. In most of the experiments undertaken a constant quantity of water was used, the variable factor being different materials under investigation of varying dry weights of samples of such materials mixed with the given weight of water.

X

I

FIG.2 Upper curve: Reference water-curve for 1.9555 grams distilled water. Lower curve: Observed volume change of 4,2836 grams toluene upon exposure to temperature change.

However, as later resrlts will indicate, it became necessary to calculate the "water curve" one might expect with any given quantity of water. Also, it was desirable t o correct the water curve shown in Fig. 2 in certain instances when relatively large quantities of solid materials were present in the system. I n order to make such calculations, it was assumed that the slope of the toluene-ice curve was determined almost entirely by the amount of toluene which was present. Therefore the quantity of toluene remaining in the dilatometer after the expansion produced by freezing would be directly related to the contraction which would result from lowering the temperature of the system. Fig. z shows the agreement which was found between the calculated water curve and the measured water curve, the calculated curve coinciding almost exactly with the experimental curve. Therefore, in later experiments a theoretical water curve was utilized, and in certain experiments corrections were made for the volume of the toluene displaced by the dry sample added to

FREE AND BOUND WATER

1s GELS

397

the toluene-water system. Where these corrections were made, the value taken for the thermal expansion of toluene was that observed in these studies, rather than the theoretical coefficient of expansion of toluene.25 InTable I1 are given the readings fromwhich the experimental water curve and the theoretical water curve were constructed.

TABLE I Volume Change of Toluene with Temperature Change Observed Temperature "C

-

Contraction of Capillary Toluene Column cm.

0.0

0.00

7 .O -19.0 -26.0 -31.0 -23.1

3.65 10.50

14.40 17.2j

12.80 6.60

-12.1

TABLE I1 The Volume Change of the System Toluene-water with Temperature Change (1.9555 grams water, 4.2838 grams toluene) Treatment and Process

1st freezing

Experimental Water Curve Length of Temperature Capillary Toluene Column "C cm. 0.0

-11.1 -17.1 -23.0

-29.0 -31.0 -34.1 -40.6 -48.0 -43.7 -39.9 -37.5 -32.0 -18.0 - 7.0

and freezing

0.00

+I7.90 +16.10 +14.20 $12.35 +11.70

+IO.80 8.75 6.30 7.70 8.95 9.75 +II.35 +I5.75 +I9.25

+ + + +

+

0.0

0.00

-16.1 -11.1 - 5.0

-I-16.30 +I7.85 -1-19.75

Calciilated Kater Curve Theoretical Length of Capillary Temperature Toluene Column "C cm. -10.0

+18.40

-20.0

+I5.25

-30.0 -40.0 -50.0

+I2.20 9.10 6.00

+ +

3 98

.


As already noted, the density change which results when I gram of water freezes, causes an increase in volume of o.090740 C C . * ~ I n the two determinations tabulated in Table I1 the expansion indicated a volume change of 0.094020 and 0.094025 cc. per gram of water. Again no cause can be assigned to this discrepancy. In no determination which we have made on the freezing of pure water was the theoretical value for expansion due to crystallization obtained. Accordingly, in the theoretical water curves constructed for our experiments we have taken our experimental value for the coefficient of expansion of water upon freezing. Studies on gelatin.-The gelatin used in the following studies was Difco Standardized Bacto-gelatin. The product was not further purified. The samples were of air-dry granular gelatin taken from a tightly stoppered container. I t was used air-dry, since oven-drying was found to produce a marked effect on the behavior of water-binding. The dilatometer was filled as follows: The dry, granular sample of gelatin of the desired weight was first placed in the well-dried dilatometer bulb. The desired quantity of water was then pipetted into the dilatometer bulb and toluene was immediately added so as to preclude loss of water through evaporation. Toluene was added until the bulb was practically full, a t which time the ground-glass stopper was tightly fitted into place and firmly secured. A very thin film of stop-cock lubricant was found to be of value in preventing leakage of the material in the dilatometer. By this method of filling, a small bubble of air would remain in the dilatometer, but the air was readily removed by passing a very fine wire through the capillary stem into the bulb. Then by holding the stem upright and exposing the dilatometer to small temperature fluctuations, the air was removed in a series of small bubbles. If the dilatometer contained an insufficient quantity of toluene, more was added by permitting the liquid to run down the fine wire into the capillary or bulb until the desired quantity was present. The most serious obstacle encountered in these dilatometric studies was the appearance of an appreciable quantity of air during the process of freezing. This air was not held mechanically in the system nor was it held on the surfaces of small solid particles, for it could not be removed by reduced pressure and it appeared gradually, as a result of the freezing process. Distilled water, which had been boiled and cooled under toluene, readily “froze out” relatively large bubbles of air, which indicated that the air was held in solution. It was often necessary t o repeat the freezing and thawing process from two to five times before records were taken in order to “freeze out” all of the air. The appearance of air, after freezing began, distorted the true shape of the curve, and therefore only those determinations in which no air appeared were accepted as indicating the behavior of colloidal systems upon freezing. Gelatin was found to give exactly reproducible data upon repeated freezings, which fact was accepted as evidence that the figures which are to follow represent a true picture of the freezing process.


399

I n Fig. 3 and Table I11 are shown the data obtained by freezing 2.3080 grams of pure water in the dilatometer. The points when plotted with change in length of capillary toluene column as ordinates and with temperature as the axis of abscissae (as in Fig. I ) , gave a straight line. When an approximately 10% gelatin gel containing the same weight of water was frozen in the same manner, an essentially straight line was again obtained. It will be noted that the readings taken during the warming of the frozen material fell, within experimental error, on the line drawn through the points given by the temperaturelowering process.

The dilatometric behavior of a gelatin gel upon exposure to lox temperatures.

Conclusions drawn by other workers *,'* 36 56 indicated that cooling the system gradually to lower and lower temperatures should cause a gradual and progressive freezing of the water in the system and that the resulting curve would resemble in general form the shape of vapor pressure curves obtained when similar materials are desiccated. Fig. 3 shows that, almost within the limits of experimental error, the points of the temperature-raising curve fall on those of the temperature-lowering curve. Also that practically no change was found in the quantity of water which will freeze beyond the point a t which initial freezing took place (- 8.0"C) even though the temperature was dropped to - 48.6OC. Table I V and Fig. 4 show the effect of repeated freezing upon a gelatin gel. The gelatin curve is the result of two consecutive freezings, the first temperature lowering being to - 48. 6OC, the system then being raised to - 6.0°C, and again lowered to approximately -60.0°C. Agreement between


400

TABLE I11 Showing the Data obtained by freezing 2.3080 grams of Distilled Water and a 10.06% . - Gelatin Gel containing 2.3080 grams of Water Distilled Water Observed I angth of Temperature Capillary Toluene Column

"C

Gelatin Gel Ohserved Temperature

"C 0.0

0.0

- 2.4 - 6.0 - 9.1 - 9.1

Length of Capillary Toluene Column cm.

-

2.0

6.0 7.9 7.9 9.1

0.0

-

0.45 I .30 - 1.70 +20.50

-10.1

-

-16.1

-11.1

+ZO.OO

-21.0

-13.1 -17.1 -19.0

+I9.50 +18.60 +18.00 fI7.30 +16.50 +I5.50 +I4.50 +14.00 +I2.80

-29.3 -34.2 -41.7 -46.6 -49.0 -45.9 -40.7 -35.8 -26.0 -20.0

- 14. I -12.1 7.0 3.0

-

-22.2

-25.5

-29.5 -33.7 -35.4 -40.3 -43.0 -44.4 -48.6 -45.6 -42.2 -40.7 -35.0 -29.2 -26.7 -21.5 -19.7

+20.30

+I2.00

$11.80 +10.60 f11.30 4-12.60 +13. IO +14.20 +I5.00

+16.10 f16.90 +17.90

-11.1

+20. I O

-10.1

+20.40 +20.60

-

-

-

9.1 8.0 7.0

+20.80

+20.go

the points representing each determination could hardly be closer, even though in the second case the temperature decrease was 20% greater than in the first. Particular attention is called to the fact that as shown by Fig. 4 the temperature-lowering and temperature-raising processes do not result in coincident lines. A consideration of this fact is not imperative at this time; rather, it is important that each line is the result of plotting the closely agreeing readings from two separate determinations.

FREE AND BOUND WATER I N GELS

4a1

In studying a system as concentrated as that reported in Table IV, it was found necessary in order to obtain a homogeneous gel to immerse the dilatometer in warm water and convert the mixture into a sol, which then was allowed to set to a gel. Accordingly, it was deemed advisable to ascertain whether or not the temperature at which peptization took place had any effect upon the subsequent freezing behavior of the system. Table V and Fig. 5

The effect of repeated freezing upon the dilatometric behavior of a gelatin gel.

I

5

I

I O

-8s nmprrrr"r0,

.a --c

-35

.M

..,

.I

FIG.5 The effect of the temperature of hydration upon the dilatometric behavior of a dilute gelatin gel upon exposure to low temperature.

IVAN D. JONES AND ROSS AIREN GORTNER

402

TABLE IV Showing the Data obtained by freezing 0.9300gram of Distilled Water and a soyc Gelatin Gel containing 0.9300 gram of Water. .. ~

Distilled Water Observed Length of Temperature Capillar Toluene Coyumn "C cm. 0.0

- 5.0 - 7.5

-

8.2

0.00

Treatment and Process

0.4650 gm. gelatin Observed Temperature

"C

1st freezing

0.0

- 4.0 - 6.0 - 8.0

-2.00

-3.05 -3.35 -3.70

+ 0.9300 pm. HzO Length of Capillary Toluene Column cm. 0.00

-1. j j -2.25

-10.1

-4.00

-10.1

-10.1

$6.00 +5.05 +3 .85 +3.25

-14.1

-3.05 +3.85 +3.40 + 2 .oo

-17.1

+I.OO

-4.25

-1.40

-31.0 -34.1 -41.4 -48.6

-2.50

-41.4

-3.40

-34.1

-5.25

-4.40

-21.0

-0.20

-5.30 -6.30 -8.50 -9.65 -7.90 -3.60

-11.1

+3.60

- 9.1 - 8.0

+4.20

9.0

-12.4 -15.6 -17.1

-21.8 -28.2 -31 .o -33.2 -35.8 -38.3 -40.7 -45.9 -48.2 -44.4 -34.1 -26.0 -16.1 6.0

-

-

+I.25

-0.50

+3.65 +7.55

-

8.0

7.0

- 6.0

Thawed and re-frozen 2nd freezing

*(approximated)

-5.35 -8.05 -11.05

-8.10

+4.55 +4.80 +4.95

- 6.0 - 8.0 - 8.0

$3.85

-17.1

+0.85

-24.0

-1.70

-32.0

-4.65

-41.4 -47 '1 -60.0* -45.9 -38.3 -28.0

-2.45 -3.20

-7.65

-10.60 - 14.25 -9.95 -6.85 -2.95

-23.0

-1.05

-18.0 -13.1 - 9.1 - 8.0

+0.85 +2.70

+4.05 $4.65


403

show data resulting from such a study. I n Fig. 5 the water curve for 0.9300 gram water from Fig. 4 has been drawn as a reference line. I n this experiment 0.1000grams of dry gelatin was added to 0.9300 grams of cold water, and the gelatin was allowed to swell for eight hours. The dotted line in Fig. 5 represents the curve obtained upon freezing the mixture. The gelatin curve gradually approaches the water curve, actually crossing it a t the very low temperatures. This crossing may be explained by the fact that the water curve has not been corrected for the presence of the dry sample. It should be noted that there is no indication of gel alteration, since temperature-lowering and temperature-raising processes result in an identical line.

+LI

-/S I

-,o

-,,

.M

. . ? ,

du

4,

-10

-4s

'

Trmpratun. -.C

FIQ.6 The rate of establishment of equilibrium of ice formation in a 50% gelatin gel at different low temperatures with reference to the effect of elapsed time upon the subsequent behavior of the gelatin gel upon freezing.

The lower line of Fig. 5 is the plot of the data resulting from two consecutive freezings of the system described above after the hydrated gelatin has been converted to a sol by warming a t a temperature of 6ooC. Again, it is considered that no alteration was effected by freezing, as most of the readings from the two freezing processes lie on the same straight line. It. is evident from an inspection of the data in Table V and the graphs in Fig. 5 that the temperature a t which gelatin is peptized plays a r81e in determining the water-binding capacity of a gelatin-gel as measured by dilatometric methods. The difference in water binding probably is the result of a more complete hydration of the ultimate gelatin particles when peptized a t the higher temperatures. One may picture the gelatin hydrated a t room temperatures as consisting of particles enveloped by thick shells of adsorbed water, with many particles still retaining their individuality as complebely as

404



405

Table VI presents data from a study of the effect of aging upon the behavior of a gelatin gel when frozen and also a study of the length of time which is required to establish an equilibrium in the freezing process. A part of the data in Table VI is graphed in Fig. 6. The system was set up by adding 0.4650 grams of dry gelatin to 0.9300 grams of water and permitting it to swell a t room temperature for three or four hours. The first freezing study is represented in Fig. 6 by the black dots (temperature-lowering) and by crosses (temperature-raising). A second freezing study was made after this system had stood at room temperature for 1 5 hours. The results of the second freezing are not plotted on the graph, but all points fell on the previously determined line. The system was then permitted to stand for five days a t room temperature and when re-frozen at - I I . I O C gave the same dilatometer reading as was obtained at this temperature for the first freezing. The system was now warmed in a water bath to 5ooC for 20 minutes, resulting in the formation of a very viscous sol which upon cooling set to a gel. After standing a t room temperature for 2 5 hours the system was again frozen. The temperature-lowering readings are represented in Fig 6 by the black circles and the temperature-raising curve by a cross within a circle. I t will be seen that equilibrium was apparently established at a much lower value of indicated free-water at temperatures from -6°C to -23OC than in the previous determinations. At temperatures lower than -23OC the readings all fell on the previously observed line. The temperature-raising curve coincides with the values previously obtained for the temperatureraising process. I n this instance an apparently decided effect was produced by the initiation of the crystallization process, as the readings between - 6°C and - 23OC differ so greatly in the temperature-lowering and temperature-raising processes. With more dilute gelatin gels, Figs. 3 and 5 , the temperature-raising and temperature-lowering curves coincided. Therefore it was assumed that some factor accompanying the increased concentration of the gelatin must be responsible for the lack of coincidence of the two curves in Fig. 6. A pronounced retardation of the velocity of ice crystal formation resulted when any system in which material was dispersed in water was frozen. It was also apparent that the greater the concentration of the dispersed material, the greater was this retardation effect. This retardation of velocity of ice crystal formation has been observed with molecularly dispersed substances by Walton and Brann.B1 The velocity changes observed in our study have been similar to those reported by Callow' in a detailed study of the rate of ice crystal growth in super-cooled gelatin gels. Therefore it seemed probable that as the point of equilibrium was approached, crystal formation proceeded a t such a slow rate that the small changes were not measurable in the time which we selected for observation. I n order to test this point, the system previously studied was thawed and re-frozen by holding it for a period of two hours at - 18OC. The reading indicated that equilibrium had been established, as was shown by comparison

406


TABLE VI A Study of the Effect of Time, Temperature and Repeated Freezings upon the Behavior of a 50% Gelatin Gel and upon the Establishment of Equilibrium in a Gelatin System-using Dilatometric Technic. Gelatin 0.4650 gram, Water 0.9300 gram Treatment and Process

First freezing Gelatin swollen a t room temperature 3-4 hrs.

Observed Temperat we 0.0

- 4.o - 6.0 - 6.0 - 7.0 - 8.0 - 9.1 -11.1

-16.1 -24.0 -34.1 -29.0 -40.7 -47. I -43.6 -36.7 -31.0

-19.0 -13.1 -11.1

-

Above sample thawed and kept a t room temperature for I 5 hours Re-frozen 2nd freezing

9.1 7.0

Length of Capillary Toluene Column 0.00

-1.45 -2.25

+4.25 $4.10 $3.90 +3.70 $3. I 5 $1.55 - I . 25 -4.95 -3.15 -7.45 -10.15

-8.65 -5.95 -3,85 $0.75 $2.95 $3.60 $4.20 $4.70

0.0

-

-

6.0

7.0 7.0

9.1

-11.1

-17.1

-33.2

-24.0 -17.1 -12.1 -10.1

-

8.0

$3.85 44.40


407

TABLE VI (Continued) A Study of the Effect of Time, Temperature and Repeated Freezings upon the Behavior of a, 50% Gelatin Gel and upon the Establishment of Equilibrium in a Gelatin System-using Dilatometric Technic. Gelatin 0.4650 gram, Water 0.9300 gram Treatment and Process

Above sample thawed and kept a t room temperature for 5 days 3rd freezing Above sample thawed and held a t 50.0°C for 2 0 minutes Stood a t room temperature for 2 5 hours Re-frozen 4th freezing

Sample from 4th freezing thawed and re-frozen 5th freezing

Observed Temperature 0.0

-11.1

0.0

- 5.0 - 6.0 - 6.0 - 7.0 - 8.0 - 9.1

Length of Capillary Toluene Column 0.00

+3.15 0.00 -1.90 -2.20

$3.00 +2.

70

+2.45 + 2 . 20

-11.1

+I.70

-17.1 -23.0

-0.35 -0.90

-28.0

-2.70

-32 . o

-4.20

-39.9

-7.05

-33.2 -24.0

-4.45

-18.0

+I.25

-14.1

+2.70

-

f4.25

9.1 0.0

-

-

7.0

-1.05

0.00 -2.60

8.0

-3.00

8.0

$2.55

-11.1 -15.2

$1.90 +I.20

-18.0

+0.20

-21.0

-0.80

-23.0

-1.45

-27.0

-2.85

-31.0

-4.35

-31.0 -41.4

-4.25 -8.20

-35.8

-6.00

-26.0

-2.30

-16.1

+1.60

-10.1

+3.65

408


TABLE VI (Continued) A Study of the Effect of Time, Temperature and Repeated Freezings upon the Behavior of a 50% Gelatin Gel and upon the Establishment of Equilibrium in a Gelatin System-using Dilatometric Technic. Gelatin 0.4650 gram, Water 0.9300 gram Treatment and Process

Above sample thawed at room temperature and refrozen 6th freezing Held at - I 5 . oo to - 18. o°C overnight Above sample melted a t jo.o°C and re-frozen. Held 2 hours at - 18OC 7th freezing

Above sample melted at room temperature Frozen and held I hour a t - 2 1 O c 8th freezing Above sample held at room temperature overnight To test time factor for equilibrium 9th freezing Frozen and held a t - I I O C for: I . j hours 3.25 " 5.25 ,' 15.50 'I Above sample cooled gradually from - I I .o°C to - 18,o"C over period of 7 hours Held at - 19. o°C 13 hours 10th freezing

Observed Temperature

Length of Capillary Toluene Column

-19.0

-0.60

0.0

- 19.5 -15.2 -12.1 - 9.1

-21.0 -16.1 -11.1 - 9.1

-11.1

-11.1 -11.1

-11.1 -20.0

-15.2 -11.1 - 9.1

0.00

-0.50 -2.25

-3.30 -4.05 - 0 . IO

i-1.85 +3.50 +4.05

not in equilibrium +2.35 +2.60 +2.60 0.00

+1.90 f3.30 +3.95

before they were hydrated, whereas if the somewhat hydrated mixture of gelatin and water is warmed to the point a t which a sol is formed and is then permitted to cool, a jelly results which from all visible appearance is homogeneous. Presumably the particles in this gel present a greater area of interface to the water phase than does the system where only swelling has taken place.


409

with the previously determined values for -1g.5'C, after the system had been subjected to a temperature as low as -40°C. I n a repetition of the thawing and freezing process a period of I hour at - 2 IOC was found to give an equilibrium reading. However, on re-thawing and re-freezing a t a temperature of - I I , IOC, equilibrium was established only after a lapse of more than three hours. These data are taken as evidence that the great difference observed between the temperature-lowering and -raising processes in the particular case represented in Fig. 6 (Table VI-4th and 5th freezings) is partially the result of lack of attainment of equilibrium in the crystallization process. This may further explain the almost horizontal shape of the temperaturelowering curve (black dots) between the temperatures -6°C and -IIOC in Fig. 6 and the similar flattening of the temperature-lowering curve of Fig. 4 between the temperatures - 8.0°C to - ~o.oo"C. It is probably also the reason why in both Figs. 4 and 6 the temperature-raising curve lies slightly above the temperature-lowering curve. This difference in both cases is more marked a t the temperatures nearer o°C, which is the temperature range at which estab!ishment of equilibrium would be slower. Undoubtedly another factor affects the behaviour of gelatin gels upon freezing. This is indicated by the dotted line in Fig. 6, but has been observed in other experiments. I n all of our dilatometer studies we have observed that the sols, gels, or even pure water would readily undercool to a temperature of from -6.0°C to -9.oOC before crystallization of ice began. It is possible that the size of the dilatometer and the relatively immobile condition in which the liquid was held was responsible for this great under-cooling. I n one experiment, when the dilatometer was being standardized with a given quantity of distilled water, the water sample was broken into two nearly equal-sized globules, well separated from each other by a layer of toluene. Upon freezing, one globule froze a t - 8.0"C and the system was held at this temperature until equilibrium was established. This equilibrium represented the freezing of only one globule of water. The temperature was then slowly lowered to - I I .o°C, the contraction in the system continuing a t a uniform rate. The second globule of water froze sharply a t - I I.o"C. I n this instance, the only factor involved in the establishment of a second equilibrium was the second initiation of ice crystallization. Obviously the earlier apparent equilibrium between 8.0 and - I I .o°C was spurious. The same effect is evident in the freezing of the gelatin gel represented by the dotted line in Fig. 6 . Freezing began a t - 6.0°C. It progressed to a point a t which about soyoof the water waa frozen. Then for some unknown reason crystallization ceased and contraction set in with decreasing temperature. This contraction continued at a rate which indicated that no additional quantity of water was being frozen. At - I 7.0"C crystallization began again and a true equilibrium was established at -23.0"C. This fact was even more evident to the experimenter than is indicated by the graph.

-

I V A N D. JONES AND ROSS AIKEN GORTNER

410

In order that this interrupted freezing may occur, it is not necessary to have an actual division of the sample, providing the portions are separated by some non-freezing substance. With the gelatin sample involved, the system had been warmed and mixed until it was apparently uniform. The gelatin gel, resulting from cooling, formed a continuous layer across the bottom of the dilatometer. Hardy,*l in microscopical studies of freezing in gelatin gels, observed that the freezing process was intermittent in certain of his gels. He found that “when a pause occurs freezing starts again, not a t the original face but a t a new face within the gel, thus leaving the characteristic membrane of dehydrated gel behind.”

,e

8

The “intermittent freezing process” postulated by Hardy and the observed interrupted freezing noted above may be due to different causes. However, the evidence presented by Hardy demonstrates the formation of an actual boundary between a freezing portion and yet unfrozen portions of the same sample. Hardy also pointed out that in certain phases of intermittent freezing the process occurs “too slowly t o be followed.” Thus it may be that the intermittent process proceeds at an irregular or interrupted rate. The significance of an interrupted freezing process in elastic gels cannot be overlooked when one is studying frozen samples by physico-chemical technic. It must also be considered in determining the temperature at which a sample should be frozen. Fig. 6 shows that maximum freezing occurs a t a temperature not lower than -II.I’C if sufficient time ( 5 hrs.) is given for the establishment of


41 1

equilibrium. Equilibrium is attained a t -21.o'C in less than I hour, but there is no evidence that more water crystallized at - z I .o°C than a t - I I. I'C, once equilibrium is attained. A few experiments were made involving the behavior of gelatin gels of varying concentrations during the process of freezing. The data are presented in Table VII. Fig. 7 is a graphical presentation of the data obtained from the most dilute concentration (2%) and the highest concentration ( 3 2 % ) .

COnSe"l?aho"

or p n x

FIG.8 Showing the weight of water bound in gelatin gels of different concentrations a t the temperatures of - 10.o"C and -3o.oOC. Expressed as gram water bound per gram dry gelatin.

In these studies 1.955 grams of water and the desired weight of gelatin were mixed and warmed to the point a t which solation was complete. With the 2 % sol no correction was made for the effect of dry gelatin in the system. It will be seen that at -6.0°C, the point of initialfreezing, theindicated boundwater content is 9.357&. At a temperature of -44.4'C,the bound-water content is, within experimental error of the method, exactly the same (9.35700). It is also shown that one is justified in considering the temperature-lowering and temperature-raising curve as a straight line. The same may be said of the curve representing the freezing of the 32y0 gel. The behavior was observed between the temperatures of -8.o"C and -32.ooC, and all points fall easily on a straight line. If now the observed


412

TABLE VI1 Data showing the Effect of Freezing upon Gelatin Gels of Different Concentrations- using Dilatometric Technic 2 % gelatin gel 16% gelatin gel .0391 grms. gelatin 1.955grms. water .3128grms. gelatin 1.955 grms. water

Observed Temperature "C 0.0

-

4.0 6.0 9.1

-12.1

-21.1 -33.2 -38.3 -44.4 -40.6 -31.0 -22.0

-18.1

-

9.1

Length of Ca illary Toluene Corumn cm.


"C

0.00

-

I .oo

+I7.40 +16.55 +15.60 +I2.80

+ + 7.40

9.10

+ 5.40 + 6.75 + 9.75

0.0

-11.1 -25.0

Length of Ca illary Toluene Coyumn cm. 0.00

$14.70 S10.70 8.70

-47.2 -38.3

+ + 6.50 f 4.30 + 6.90

-24.0

+II.OO

-16.1

+I3.35 +15.35

-32 . o -39.9

-

9.1

+I2.50 +13.80 +16.60

8% gelatin gel 32y0 gelatin gel 0.1564 grms.ge1atin 1.955 grms. water 0.6256 grms.gelatin 1.955 grms. water Observed Temperature "C 0.0

Length of Capillary Toluene Column cm.

Observed Temterature C

0.00

-

0.0

Length of Capillary ToIuene Column cm. 0.00

- 4.0 - 7.0

+16.10

-16.1

$12.85

- 9.1

+15.55

-22.0

+II.20

-11.1

+I5.00

-29.1

-13.1

+14.40

-33.2

-21.0

+I2.00

-22.0

++II.ZO 8.05

-33.2 -39.9

-48.0 -36.7

I .IO

+ 8.45 + 6.50 + 3.90 + 7.35

Thawed and re-frozen 0.0

-17.1 -13.1 - 9.1

0.00

+I3.15 +14.35 +15.55

-14.1

-11.1

-

8.1

+13.35

+ 9.15 +14.30 +14.90


413

water curve be corrected for the presence of the dry gelatin (as indicated in Fig. 7 by the dotted line), the percentage of bound-water present at -8.ooC is 20.56 and a t -33.Ooc is 22.43. This most concentrated gelatin gel curve, then, may be considered as practically a straight line parallel to the theoretical water curve. The expression of the percentage of bound-water a t the temperatures - IO.OOCand - 3o.o0Cgives an indication of the degree of parallelism between the observed gelatin curve and the theoretical water curve. Table VI11 shows an actual slight decrease in total free-water content when the temperature is lowered from - IO.OOCto - jo.o°C. This is the reverse of what would be expected if bound water were converted into ice by a lowered temperature. Corrections for volume of toluene displaced by the sample were based on the density of dry gelatin. Thus, knowing the weight of the dry gelatin used in the determination, the volume of toluene displaced by the gelatin could be calculated. The density of the gelatin was determined experimentally and was found to be 1.385. This method of correction is subject to question, for it is known that the volume occupied by gelatin after being wetted by water is actually less than the sum of the volume of the dry gelatin plus the volume of the water taken. Svedbergs’ found the contraction caused by wetting gelatin to be in the neighborhood of 0.055 cc. per gram of gelatin. This value is so low that, this could not be the source of appreciable error under the conditions of our experiments. His study does justify the question, however, as to what may be the actual volume of the “frozen-out” gelatin. Table VI11 and Fig. 8 show the data for bound water in gelatin gels a t - 10°C to -3ooC, as calculated from the dilatometer readings recorded inTable VII. The calculations have been expressed in the grams of water bound per gram of dry gelatin. Freundlich16 has given a general expression for an adsorption reaction which expressed mathematically is the equation for a parabola. The logarithmic expression of a parabolic curve is a straight line. Plotting the arithmetic values of Table VIII, where the abscissa represents the grams water bound per gram dry gelatin and the ordinate represents the concentration of the gelatin gel, a smooth curve (A-Fig. 8) was obtained which appeared to be parabolic. When the logarithmic values were plotted, the resulting points fell practically on a straight line (B-Fig. 8). Accordingly it appears probable that water-binding in gelatin systems is an adsorption reaction. TABLE VI11 The Bound Water in Gelatin Gels as a Function of Gel Concentration Gel Concentration

%

Bound water expremed as per cent of total water in syetem

- 10°C

-30°C

%

%

2

9.35

8

15.10

16 32

16.16 20.56

9.35 15.19 16.82 22.43

Water bound per gram dry gelatin

- 10°C grams 4.675 I .888 I .OIO 0.643

-30°C grams 4.675 1.899 I .os1

0.701

4r 4


All of the studies which we have made on gelatin gels were made on gels prepared from essentially iso-electric gelatin (pH 4.7-4.8). Some experiments were made to study the effect of varying pH upon the dilatometric behavior of dilute gelatin gels when exposed to low temperature, but no observable difference due to a pH effect could be detected.

FIG.9 The dilatometric behavior of the thick portion of egg white upon repeated exposure to low temperature.

T h e Behavior of the Thick Portion o j Egg W h i t e f r o m Fresh Eggs upon Freezing.-A short study was made upon the thick portion of the white of fresh eggs. This material was chosen as a natural hydrophilic colloidal system which would serve to extend the range of the bound-water studies. No attempt has been made to conduct a complete study of the behavior of egg white upon freezing. Dilatometric analyses were made of two samples taken from two eggs not more than six hours old. The sample taken in each instance was a portion of the unmixed thick white of one egg. It was considered, as has been pointed out by St. that mechanical mixing of the egg white would alter the colloidal nature of the system. The experimental data obtained from this study are given in Table IX and Fig. 9. These data indicate that in each sample the freezing process was completely reversible. The readings for the temperature-raising process when plotted fell on the curve representing the temperature-lowering process. Also, the data taken during a second freezing determination fell practically on the curve representing the first freezing process.

F R E E AND BOCND WATER IN GELS

415

I n this respect the behavior of egg white was similar to that previously observed for gelatin. However, with egg white the curves obtained were not straight lines. A more surprising fact was that they receded somewhat from the theoretical water curve a t the lower temperatures. This is clearly indicated in Fig. 9 and shown numerically in Table X. The fact that the observed curve gradually receded from the theoretical curve was an indication of a contraction greater than would be expected from thermal contraction alone. If some of the “bound” water had frozen we should expect an expansion. Egg white cannot be considered a chemical entity but is a colloidal system containing egg albumin and other proteins.

TABLE IX A Dilatometric Analysis of the Effect of Freezing of the Thick White of Fresh Hen’s Eggs Weight

Sample I grams. Water 85.68%

3.1929


“C

1st freezing

Length of Capillary Toluene Column cm. 0.00

0.0

-

4.0 6.0

-

0.85 1.25

9.1

-11.1 -13.1 -17.1

-22.0

-30.0 -34.1

-43.0 -49.4 -40.7 -31 . o -16.1


6.0

-

6.0

-

-

9.1

+23.20

0.90

-11.1

$22.95

-15.2

$22.30

- 2 1 .o

+21.25

-29.0

+19.70

-34.1

+18.90

-26.0

+20.20

-16.1

+22.20


TABLE IX (continued) A Dilatometric Analysis of the Effect of Freezing of the Thick White of Fresh Hen’s Eggs Sample 2 Weight

2.3008 grams.

Water 8 8 . ~ 3 7 ~

Observed Temgerature

C

1st freezing

0.0

-

4.0 7.0 7.0 9.1

Toluene Column cm.

-

0.00

1.30 2.25

SI5.90 +15.50

-11.1

+I5.05

-14.1

+I4.50

-17.1

d-13.45

-22.0

+11.go

-35.8

+ 9.75 + + 6.65 + 3.55 + 5.4c + 7.7c

-26.0

+IO.SC

-29.0

-34.1 -39.1 -48.0 -43.0

-16.1


Length of Capillary

8.20

+13.85

9.1

+15.60

6.0

+16.10

-16.1

+13.85

-11.1

+15.20

Soren~en;~ made an extended study of crystalline egg albumin. His crystals were separated under conditions of carefully controlled concentrations of hydrogen ions and ammonium sulfate. H e found that the crystals were actually hydrous egg-albumin sulfate containing 0.22 grams water per gram dry albumin. From the fact that egg albumin may be separated in the crystalline form from a colloidal system,it seems possible that the observed greater contraction of egg white at low temperature might be due to an orientation of the hydrated egg albumin into a definite crystal lattice. This crystallization resulting from exposure to the low temperatures would take place with the formatiorl of a more closely packed space lattice and accordingly contraction would result.


417

Studies on Systems of the Non-elastic-Gel Type Studies on Actiaated Silica Gel.-Pulverized

silica gel, obtained from the Silica Gel Corporation, was sieved; the portion which passed through the 60-mesh and was held on the go-mesh sieve was utilized in this study. This portion was activated by heating in a vacuum oven at I j0"C for z hours. Systemswereset up by mixing 0.4888 gram, 1.173ograms, and 1.574 prams of activated silica gel with 1.955 grams of distilled water. Heat was liberated when the silica gel was wetted, giving evidence of a decided adsorption of

FIG.I O The dilatometric behavior of the system Activated Silica Gel-Water upon exposure to low temperature. (System consisting of 1.95 grams distilled water and 0.4888 grams Activated Gel).

hxi

water. Table XI shows the dilatometric records which were made on the z5%, 60% and 80.5% silica gel systems. The data are graphed in Figs. IO, I I and I 2 . In these figures correction has been made for the volume occupied by the dry silica gel, taking 2.20 as the density. Fig. I O shows the shape of the curve and the general agreement of experimental data obtained when the 25c0mixture was frozen to -48.ooC. The points of the temperature-lowering curve fall on a straight line which is practically parallel to the corrected water curve. The line marked with the black crosses represents the temperature-raising curve. These points also fall on a straight line, but as the temperature is raised this line slowly diverges from the temperature-lowering curve. Fig. I I is the graphical representation of the effect of freezing upon a 6 0 7 ~ activated silica gel system. The black dots on line AB represent the readings taken during the first freezing. The line BC represents the slope of the tem-

IVAN D. JONES AND ROB8 AIKEN GORTNER

418

FIG.I I The dilatometric behavior of the system Activated Silica Gel-Water upon repeated exposure to low temperature. (System consisting of 1.955 grams distilled water and 1.173 grams Activated Silica Gel).

TABLE X A Dilatometric Study of the Thick White of Fresh Hen's Eggs. The Bound Water calculated from the Expansion which was observed due to Ice Formation and the Theoretical Expansion which should have resulted if the Total Water present had frozen Temperature

"C - 6.0 -10.0

-20.0

-30.0 -40.0 -48.0

"Apparent Bound-Water" at Different Temperatures Expressed as % of Total Water Present in System Sample I Sample 2

%

gr,

17.36 16.19 14.86 13.36 13.36 14.36

20.18

18.16 17.26 17.71 18.16

19.73

perature-raising curve. Again, this is a straight line lying decidedly above the temperature-lowering curve. The sample was then thawed and a second freezing was conducted. The open circles, representing these readings fall on the previously formed line. The temperature was again raised to - 26.ooC and the two readings taken were found to fall on the corresponding line of the first temperature-raising process.

FREE AND BOtiND WATER I N GELS

n-ew".

419

-15

FIG.12 The dilatometric behavior of the system Activated Silica Gel-Water upon repeated exposure to low temperatures. (System consisting of 1.955 grams distilled water and 1,574 grams Activated Silica Gel.)

Accordingly, repeated freezings gave reproducible data,* which indicated that no alteration of gel structure had resulted from the freezing process. At this time the sample which was at a temperature of - 26.0°C was cooled to an approximated temperature of - 75OC. Readings were then made as the temperature was raised to - 2 I .o°C. Line D E represents the curve obtained. Point D lies very decidedly above point B, indicating that an appreciably greater quantity of water has been frozen at - 75.0°C than a t -48.ooC. After warming the sample to - z I.oOC (point E) it was held a t this temperature for 3 hours, A reading then taken has been represented by point F. Point G indicates the reading when the sample was warmed a t - 11.o"C. On again lowering the temperature to - 21.o"C, without further thawing the system, it was found that the equilibrium had not been disturbed. Fig. 12 represents the behavior of the 80.5% mixture upon being frozen. The first temperature lowering was to only -33.2OC. The blackdotsand black *It should be added that this statement holds when the termination of the temperaturelowering process has been a t approximate1 the same temperature. Thus practically no hysteresis could be observed between the f%st and second temperatureraising proceeaes, the termination of the first and second temperature-lowering processes being -48°C and -43.6"C respectively. Hysteresis appeared ?hen the system WBS cooled to varying low temperatures. There is a decided difference in bound water values (cf. Table XII) at the same temperatures of the temperature-raising process when the termination of the temperaturdowering process has been different. Consider as an example the 6rst and third temperature lowering processes which terminated a t -48°C and -75°C respectively. With the 80% Activated Silica Gel-Water 8 stem the termination of the first and second temperaturelowering rocems were - 3 3 . 2 4 and -49.4'C respectively. Accordingly this system likewise si?owed the effect of the lowest temperature reached in the freezing process. The differences in the bound water values represented at the different temperatures in Table XI1 are, we believe, due to the conditions mentioned above.

420

2 Y



42 1

TABLE XI (Continued) A Dilatometric Study of the Effect of Freezing upon Systems of Activated Silica Gel and Water System I11 System I1 Activated Silica water water Activated Silica Gel 1.5740 grams 1.9550 grams 1.9550grams Gel 1.1730 grams Treatment and Process

Thawed and refrozen 2nd freezing



"C

cm.

-II .I

+II.85

-19

+IO.IO

0

-26.0

-35.0 -43.6 -31 .o -26.0

-75.0 -48.0

-39.9 -31 .o -21 .o

Held 3 hours at -21.ooC

++ 8.65

Treatment and Proceas



"C

cm.

-34.1 -45.9 -A9.4 -41.4

+ + + +

6.50 5.05

4.45 6.40

6.8j

-35.8

f 5.40

+ 7.45

+

-21.0

+IO. 1 5

8.20

-11.1

+12.45

9.25

- 7.0

+12.90

+-

+ 5.80 +

7.60

+ 9.30 +II.50

-21.0 -11.1 -21.0 -11.1

crosses represent the readings taken during the temperature-lowering and temperature-raising processes, respectively. A second temperature lowering was made to 49.5OC. The readings of the second process lie on the previously obtained straight line to the temperature of -3o.o"C. At this point, as in Fig. 11, the curve makes a decided convergence toward the water curve. The temperature-raising process from the temperature -49.5OC gives a line lying considerably above that given when the sample was frozen to only -33.z°C. In Fis.3. IO, 1 1 and 12 the observed water curve for 1.9550 grams of water has been drawn. The water curve corrected for the volume of silica gel has been represented by a dotted line. Table XI1 shows the percentage of boundwater found P t different temperatures, the corrected water curve being in all cases considered to represent rooyo of free water. The behavior of the activated-silica gel-water systems is very different from that of the system, gelatin-water. These differences are (A) the temperature a t which maximum freezing occurred and (B) the behavior of the

422


frozen mass as the temperature approached the melting point of ice. In the gelatin systems maximum freezing was complete a t temperatures not lower than -6.o'C and further cooling of the frozen mass to -48.ooC was without apparent effect on the bound water content of the gelatin. With the silica gel systems gradually lowering temperatures caused increasing quantities of water to freeze; also, the temperature-raising curve lay distinctly above that for the temperature-lowering process.

TABLE XI1 Showing the Apparent Per Cent of Total Water existing as Bound Water in Systems of Activated Silica Gel-Water a t Different Low Temperatures Bound Water at Temperature Indicated Concentration of system Process -IOT - 2 o T -30°C -40% -48°C per cent per cent per cent per cent per cent per cent 10.28 9.35 9.90 10.02 25 Temp.lowering 9.81 25 Temp. raising 9.30 10.28 7.94 8.01 8.64 60 60 60

29.90 Temp,lowering Temp. raising, 1st 2 5 . 7 0 Temp. raising, 3d* -

28.50

27.57

25.00

24.30 19.16

-

Temp.lowering 38.32 37.15 Temp.raising, 1st 35.75 35.75 Temp.raising,and 31.54 30.14 * Second temperature raising'not calculated. 80 80 80

35.28 35.28 28.97

25.47 23.83 18.23

22.90

31.07

27.58 26.87

--

27.80

22.90 17.52

If the behavior of inelastic gels upon freezing is analogous to that of elastic gels as indicated by gelatin, the curves representing the temperature-lowering and raising processes should be the same, and decreasing temperatures would not increase the amount of water that could be frozen. Again, if silica gel consists of a mass possessing a structure extremely capillary in nature, as is conventionally a ~ c e p t e d , ~ ~it. can ~ ~ ,be ~ ~seen , ~ ' that the water in the gel must be held with a very great force of capillarity. This force is sufficiently great to resist the forces of crystallization when the temperature is lowered below the freezing point of water, keeping the water adsorbed in a liquid state. Finally, as the temperature is lowered the crystallization force becomes great enough to cause the capillary water to freeze. If the temperature be now raised a slight amount, it would seem that equilibrium should be established anew between the opposing forces of capillarity and crystal formation, with the result that some of the frozen water would melt and again become adsorbed capillary water. This would result in a completely reversible process, with the temperature-lowering and temperature-raising curves following the same path in spite of the fact that decreasing temperatures caused an increase in the quantity of water which could be frozen. That the freezing of silica gel systems is not a readily reversible process is indicated by Figs. IO, 11 and 12.


423

However, Figs. I I and I 2 and Table XI1 indicate that a partially reversible reaction takes place in the 60% and 80.5% samples. The slope of the temperature-raising curves in Figs. 11 and 1 2 is divergent from the corrected water curve, indicating that actually less free-water is present at - I O . O O C after the sample has been warmed from - 48.ooC than is present at the point of -48.ooC. The same tendency is more clearly indicated in Fig. 11 by the points E and F. E was the reading given when the frozen mass which had

m m ~ r w -'c e

FIG.1 3 The general dilatometric behavior of inorganic hydrogels upon exposure to low temperature, as observed by Foote and Saxton.

been cooled to -75.0°C was slowly and gradually warmed to - z I . o O C . When the temperature was held at --21.o'C for several hours, point F was found to be the measured value for that temperature, rather than E. Evidence that this is nearer the true point of equilibrium is shown by the facts that (a) after raising the temperature to - II.O'C the same value was obtained when the temperature was again lowered t o -zI.o'C and (B) when the lines DE and FG are extended they are found to intersect a t the o.o°C line. This difference between the points E and F would indicate that the velocity of equilibrium establishment is a factor determining in some degree the slope of the temperature-raising curve. Time cannot be considered the only factor, for after a period of three hours the line FG, which apparently represents the points of true equilibrium, lay distinctly above the corresponding portion of the line BC. Evidently, then, some difference was produced in the point of true equilibrium when the sample was cooled to -75.0°C rather than only to -48.ooC.

Studies on Colloidal Ferric Hydroxide Systems.-We have already noted that Foote and S a ~ t o n ~ investigated ~J~J~ the effect of freezing upon certain

424


inorganic colloidal systems. The general shape of the Foote and Saxton curve is indicated in Fig. 13. They assumed that three kinds of water existed in their gels, (A) “free-water,” (B) “apparent capillary-water,” (C) “combinedwater.” The free-water was considered to be the water which froze readily a t a constant temperature. This is indicated in Fig. 13 by the vertical line BC. As the temperature was gradually lowered to point D, the curved line C D was obtained. The gradual freezing, represented by the curved line, was considered a freezing of water held in the capillary spaces of the gel. Raising the temperature from D to near the zero point gave the line DF. D E is a straight line formed by an extension of the straight portion of line DF. DE was considered the theoretical measure of the combined water. The gradual sloping away of D F from DE was considered to be due to the remelting, as the temperature approached o.o°C, of some of the water held by capillarity. As Foote and Saxton were unable to find a point at which there appeared to be a sharp distinction between “free-water” and “apparent capillary-water” they empirically chose - 6.ooC as the temperature a t which all “free-water” was frozen. The vertical distance between the observed point C and the extrapolated point E was considered to be a measure of the “apparent capillarywater.” Our results are at variance with those of Foote and Saxton. We have shown that with gelatin, if sufficient time were given for equilibrium to be established at the point of initial freezing, the line CD was a straight line and the line D F was also a straight line coinciding with CD. With activated silica gel, lines CD and D F were both straight, though not coincident. I t seemed that two possible explanations might be advanced for the curved line obtained by Foote and Saxton, (A) that a true equilibrium was not established at the temperature of initial freezing (their point C) before the temperature was again lowered, or (B) that the sample, if not electrolyte-free, would at first freeze gradually and also begin to melt at temperatures lower than zero. The following experiments were carried out on a hydrated ferric hydroxide coagulum to clarify, if possible, the points in question. The hydrated ferric hydroxide-gel was prepared as follows: IOO cc. of 3 0 7 ~ferric chloride solution was added to 400 cc. of boiling water. The colloidal ferric hydroxide which formed was precipitated by the addition of a sufficient amount of dilute ammonium hydroxide. The coagulum was dialyzed against distilled water until no chlorides appeared in the external liquid. The coagulum was filtered and sucked as dry as possible upon a Buchner funnel. The mass of ferric hydroxide was then well mixed and aged for several days over distilled water in a closed vessel. The resulting ferric hydroxide, although having a moisture content of from 85-88%, had the consistency and appearance of thick apple butter. Tables XI11 and XIV and Fig. 14 show the data resulting from the dilatometric measurements made upon this hydrated ferric hydroxide. System I consisted of a 3.541 gram sample containing 84.22% water. Tnitial freezing took place at -6.onC, and after freezing was cornplete the contraction-expansion curve to -48.ooC was a straight line. As the temperature was gradually raised to -6.o”C there resulted a second straight


42 5

line, but as in the activated silica gel studies, the temperature-lowering and temperature-raising curves were divergent from the point of lowest temperature. Determination I1 is a repetition of determination I. I n System I1 the sample weighed 2.7071 grams containing 8 5 . 2 5 7 6 water. The solid black lines of this part of Fig. 14 represents the results of the first freezing process. I n this determination, as in the former, the points of the temperature-lowering and temperature-raising process lie on straight lines diverging from the point of lowest temperature.

FIG.14 The dilatometric behavior of hydrous ferrlc hydroxide upon exposure to low temp-Orature.

With both samples I and I1 freezing was observed to result in a very decided physical alteration. When the sample was thawed, the ferric hydroxide settled in brown flakes from clear water, the water representing a large percentage of the total volume of the sample. Thus the "apple butter" consistency of the coagulum was completely disrupted. Accordingly a second freezing was made of sample 11. Readings of this freezing are indicated by open circles and by a cross within a circle. The resulting lines follow closely the lines of the first freezing process. I n this study with ferric hydroxide the results coincide in many respects with those found by Foote and Saxton, with the exception that all of these readings fell on straight lines, while those of Foote and Saxton curve decidedly between the temperature - 10.o"C and -6.0°C. It was a matter of considerable surprise to observe that the ferric hydroxide, completely precipitated by the effects of the first freezing process, should continue to exhibit such great ability (Table XIV) to prevent water from freezing.


426

The theoretical water curves have been included in Fig. 14. I t will be seen that decreasing temperatures cause a gradual approach toward this line, as would be expected from the convergence of the temperature-lowering and -raising curves. TABLE XI11 A Dilatometric Study of the Effect of Freezing upon the State of Water in a Colloidal Ferrjc Hydroxide Coagulum Sample I. 3.541grams (water 8 4 . 2 ~ 7 ~ ) Length of Observed Capillar Remarks Temperature Toluene Cocmn “C cm.

First freezing

0.0

-

4.0 6.0 6.0 - 8.0

0.00

Sample 11. 2.7071 g r a m (water 85.257C) Length of Observed Capillary Remarks TempErature Toluene Column C

First freezing

0.0

+25.55

-

-

0.70 1.00

6.0 6.0 8.0

+25,20

-12.1

-11.1

+24,80

-17.1

-16.1

+24.05

-25.0

-22.0

$23.20

-29.0

+22.45

-33.2 -28.0

-20.0

-34.1

+21.80

-39.9

+20.95

-12.1

-47.1

f19.80

- 6.0

-41.4

+20.90

-13.1

-35.8

+21.80

-21.0

-26.0

s23.30

-25.0

-19.0

f24.45 +26. IO

- 9.1 - 6.0

+26.60

-15.2

Thawed and refrozen 2nd freezing

-12.1 -24.0

-35.0 -25.0

-16.1

Thawed and refrozen, 3rd freezing

6.0

-12.1 -23.0 -34.1 -26.0

-16.1 - 6.0

Thawed and refrozen, 4th freezing

-12.1 -22.0 -32.

o

-2j.0

-16.1

-

6.0


42 7

TABLE XIV A Calculation of the Percentage of the Total Water in Colloidal Ferric Hydroxide Systems which remained unfrozen when the System was cooled to Various Temperatures. (Data calculated from Table XIII) Sample

I

I1

Remarks

Point of Initial Freezing Tempera- Amount of tye “bound” water C per cent

Temperature of Lowest Exposure TemperaAmount of tye “bound” water C per cent

Temperaturelowering Temperatureraising

- 6

18.07

-47

‘3.63

- 6

14.85

-47

13.63

1st temperaturelowering 1st temperatureraising 2nd temperaturelowering 2nd temperatureraising

- 6

21.18

-33

17.22

- 6

19.00

-33

17.22

-12

17.02

-35

I4

-12

15.04

-35

14.25

25

Discussion The Methods which were employed.-In our studies we have calculated a theoretical water curve, and in this calculation have ignored any contraction of ice upon sub,ecting it to decreasing temperatures. This was done because it was possible to so nearly reconstruct the actually observed curve representing the freezing of 1.9550 grams of pure water (Fig. z ) , without a consideration of the thermal expansion of ice. If these water curves are erroneous, then the conclusions which have been drawn regarding the relation of observed volume increase to theoretical volume increase are in error. However the disregarding of any slight volume change of ice with temperature change has no influence upon the further consideration given to the studies on gelatin and activated silica gel, by the establishment of a “corrected water curve.” This correction has been made by allowing for the volume of toluene displaced by the mass of the dry solid used in preparing the colloidal system. Obviously, with less toluene in the experimental system toluene-colloid-water than in the reference system toluene-water, the difference in quantity of toluene present must be considered, if the expansion or contraction of the volumes of the two systems are comparable when exposed to temperature change. The Effect produced by Freezzng upon Systems of the Elastac-Gel Type-(A) Gelatzn Gels.-Under the experimental conditions of the present investigation, the freezing of the system gelatin-water was completely reversible. These results are a t variance with those reported by Molisch,36 and by Fischer and Bobertag.l* They are also in opposition to the theoretical conclusions of

428


Iiuhn,z9and Stiles,S6who considered that freezing would always be accompanied by a partial and a more or less permanent alteration of the original gel structure. The present study is in agreement with the results of Liesegang,30and M ~ r a n . ~ ~ StilesS6has emphasized the importance of the rate of freezing of colloidal systems in relation to the effects produced by the freezing process. He found that with thin sheets of hydrated gelatin the water-loss upon thawing was from 4 to 6 times greater from samples which had been slowly frozen, than from similar samples which had been frozen rapidly. Moran’s3? studies, demonstrating internal and external centers of ice crystal formation in gelatin gels, may be considered a verification and explanation of the results reported by Stiles. With the slowly frozen samples the ice formation was to a large extent external and as a consequence re-adsorption of the water was slow. On the other hand, Hardyz1observed that when disseminated ice crystals formed wzthin the gel, and when crystals of a solid solution of gelatin formed, the water resulting from thawing the frozen mass was immediately re-adsorbed. Our experimental conditions favored rapid freezing of the gels. No permanent alteration of the gel has been observed, nor have studies under experimental conditions favoring slow freezing been made. It should be stressed that when we refer to “an alteration of gel structure” it is with reference to a change in the gel structure of a behavior of the gel upon exposure to changing temperatures which is detectable and measurable by dilatometric technic. Thus in our system the gelatin gels occupied the same volume at any given temperature above 0°C after the sample had been frozen as had been occupied before freezing. 1 1 0 r a n ~believed ~ that he could detect a slight volume change brought about by the process of freezing and thawing. Also, in our experiments the process of repeated freezing and thawing gelatin gels in no way altered the quantity of water which crystallized at the initial freezing point of the system. This has been interpreted as indicating that ice-crystallization took place relatively rapidly and that the centers of crystal formation were principally internal. These results are considered as additional evidence in favor of the theory advanced in part by Hardy and in part by Fischer, Hardy postulating that dehydration could be considered a reversible process if a gel or sol resulted normally from the addition of a colloidal substance to water, and Fischer considering the process of freezing only a certain type of dehydration. Moran, by means of the dilatometer, found that, with a 43.776 gelatin gel held at - I I .o°C ice crystal formation was only complete after a period of twenty-six days. I n view of his study it is pertinent to ask whether our results represent a true equilibrium of the freezing process. Our data designed to test this question indicate that our values cannot be greatly in error. The fact that it was possible to obtain results clearly demonstrating a reversible reaction with the elastic gels, is evidence that equilibrium must have been practically complete. Otherwise, a reading of the dilatometer a t - I 5.o”C would have been distinctly different after exposure of the sample to a temperature of - 50.0°C than before exposure to this lowered temperature.


42 9

We have already noted that our dilatometric data indicate that waterbinding in gelatin gels is an adsorption reaction. Newton and GortneF and Gortner" interpret water-binding in colloidal systems of gum acacia as an adsorption reaction, although this may not hold for all samples of gum acacia [cf. Newton and Martin"]. Further evidence of the nature of water-binding in hydrophilic colloidal systems is given by Newton and Martin.44 Their gelatin data have been recalculated and are shown in Table XV and Fig. I 5. 1

'I

I

.a

I FIG.1 5 Showing the weight of water bound by gelatin sols of different concentrations. Data expressed as grams water bound per gram dry gelatin (cryoscopic data of Newton and Martin).

TABLE XV Bound Water in Gelatin Sols as a Function of the Concentration of the Gelatin Sol. Determinations by the Newton-Gortner Cryoscopic Method. Data of Newton and Martin44 Concentration of Gelatin in System per cent 0 93 I .86

Bound-Water per cent 1 .9 3.2

Water Bound per Gram of Dry Gelatin grams 2.05

I .7 0

2.77

3.7

1.31

3.66 4.55 5.43

4.4

1.17

5.0

I .04

5.5

0.96

430


Newton and Martin’s studies were carried out by the Newton-Gortner cryoscopic method, which can be used only with dilute gelatin sols, rather than gels. I t will be seen that these data indicate again an adsorption reaction due to the fact that when the arithmetic values of gelatin concentration and bound water are plotted, Curve A, apparently a parabola, resulted, whereas when the logarithmic values were plotted a straight line relationship was indicated (Curve B). We have already (Fig. 8) shown a similar relationship from our dilatometric data. A review of the literature on studies made upon aqueous gelatin-containing systems reveals a most interesting array of values for the actual quantity of water which is combined with gelatin or which exists in a ‘(bound” condition. The method of measurement of water-binding by gelatin has been different in nearly every study reported. Svedberg,j7 from studies upon the contraction of gelatin when wetted, considered the weight of water bound per gram dry gelatin to be in the neighborhood of 0.08 gram; Fischer, as cited by Moran,37 from studies on the rate of drying found 0.1gram water to be bound per gram dry gelatin. TaffeP considered the work of Sheppard and Sweet,54on the setting points of gels, as a basis for the value of 0.3 gram bound water per gram dry gelatin. from his dilatometric studies, found the value of 0.56 gram bound water per gram dry gelatin for gels of 4 3 . 7 % and 52.1% concentration. Newton and Martin’s results were obtained in a study of the freezing point depression of theoretically molar sucrose solutions. With the technic of measurement of bound-water differing so radically, it is not surprising that different values were obtained for the quantity of water bound by a gram of gelatin. Furthermore, the concentration of the gelatin in the system is a factor which influences the value for weight of water bound per gram dry gelatin. This is clearly indicated by the results of the present study (see Table VI11 and Fig. 8), and by the data of Newton and Martin (Table XV and Fig. Is). Our own data show that, in a system containing 2% gelatin a t - IO.OOC,4.7 grams of water is “bound” per gram dry gelatin, while in a 3 2 % gelatin gel at - 10.o”C only 0.64 gram of water per gram dry gelatin appears to be “bound.” This last value does not differ greatly from Moran’s value which was also obtained by dilatometric studies. From the standpoint that water-binding in gelatin gels appears to be an adsorption reaction, it was considered probable that the temperature at which the system would be frozen would influence the value for the quantity of water which was bound. However, no indication that temperature is a factor can be detected from an inspection of Figs. 3, 5 and 7, providing that sufficient time has elapsed for the establishment of an equilibrium in the freezing process. Over the temperature range which was studied, all of the water which could be frozen crystallized at the temperature of initial freezing and exposure of the system to decidedly lower temperatures was without effect upon the quantity of water which appeared to remain in the unfrozen condition. A possible explanation may be that two factors are affecting the equilibrium (A) the vapor pressure of ice decreasing with lowering temperature and (B) an adsorption “pressure” increaszng with lowering temperature. If the


43 1

increased adsorption tendency should be sufficient to just balance the lowered vapor pressure of the ice, then the amount of “bound” water (adsorbed water) should be a constant irrespective of temperature. (B) Systems of Fresh Egg Whzte.-With egg white, as with gelatin, the freezing process appeared to be completely reversible. St. Job@ made a study of bound-water in the thick portion of egg white, employing the method of RubneIS* as modified by ThoenesG0and R~binson.~gHe found that at - 5.o”C 85% of the water present appeared as bound-water, while a t - 15.o”C only 26y0 of the total water in the system appeared to be “bound.” Our values on a similar system are given in Table X. It is of interest to note that while St. John’s value a t - 5°C is greatly in excess of our values at -6°C (17.36% and zo.18oJ~)there is much less discrepancy at -15”c. Newton and Martin44 have also studied the question of bound water in fresh egg white, using a composite sample from 16 eggs. I n their studies the whites were beaten, allowed to settle for 3.5 hours, the froth was removed, and the clear liquid underneath taken for study. This clear liquid portion was then diluted to suitable concentrations. Bound water was measured by the Newton-Gortner cryoscopic method. They found much lower values than either ourselves or St. John, L e . , 0.7% bound water for a 2.37% sol, 2.3% for a 7.11% sol, and 6.5% for an I I .85% (undiluted) sol. These values (recalculated) indicate respectively 0.29, 0.30, and 0.48 gram of bound water per gram dry albumin, whereas using the dilatometric method we find (sample I1at - 6°C) 1.55 gramsof bound water per gram dry egg white. These differences in the magnitude of the values obtained by Newton and Martin and by ourselves are too great to be explained wholly on the basis of the mechanical manipulation of the sample or from the fact that their studies were made on a mixed sample of the thin and thick portions of the white. Unquestionably the method by which bound water is determined is the primary factor. Apparently a very considerable fraction of the “bound” water as measured by dilatometric and (probably calorimetric) technics can act as a solvent for sucrose and thus lead to low values when the Newton-Gortner cryoscopic method is employed. Further studies of this colloidal system must be made before final conclusions can be drawn. The Efect produced by freezing Systems of thelnelastic-Gel Type.-Pulverized activated silica gel proved to be very desirable material for bound-water studies. It was sufficiently granular to wet readily, was very stable in nature, and in repeated and consecutive determinations gave readily duplicable results, Thus, the water “frozen out” by temperatures decidedly lower than the initial freezing point was completely re-adsorbed when the sample was thawed. This ability to return after the freezing cycle to the condition existing before freezing occurred was not observed by us in other inelastic-gel systems. At no point in the recorded temperature study did we observe a temperature below which no more water would freeze if the temperature were lowered further. Possibly all the water added to these samples would freeze if sufficiently low temperatures were employed. If this be true, then water in the silica gel system may exist only as “free-water” and “capillary-water.” All the water

432

I V A N D. J O N E S AND ROSS AIKEN GORTNER

which failed to freeze at the point of initial freezing might be classed as “apparent capillary-water’’ nt this temperature. The true measure of “capillarywater” would then be all the water which did not freeze at theinitialfreezing point of the system. This point in the present studies appears to lie between - I.o”C and -2.o”C. I t is apparent, however, from Table XI1 that the system would have to be exposed to extremely low temperatures in order to freeze all of the water, since the “bound water” ( 2 . e . water not frozen) values decrease very slowly with lowering temperature. Thus in the 6oC: gel the “bound water” has decreased only from ~ 9 . 9 to 7 ~2 2 . 9 7 , in the range -10°C to -48OC and in the 80% gel from 3 8 . 3 ~ to 7 ~2 7 . 5 8 % in the same temperature range. An extrapolation of the curves for “bound water” indicates that all of the water would not be frozen until the sample had been exposed to temperatures below - 80°C. Ferric hydroxide gels exhibited a behavior similar to that of silica gel. A certain portion of the total water content froze at a constant temperature and the remaining quantity of unfrozen water gradually decreased as the temperature was lowered. Although the freshly prepared ferric hydroxide system possessed a marked degree of rigidity, the structure was readily altered by freezing. The actual affinity for water, in the ferric hydroxide coagulum, seemed to be reduced but little by freezing, but the dispersion of the particles was permanently altered. That this alteration can be the result of the disruption of capillary spaces seems unsupported by the observed behavior. If the capillaries had been so completely injured by the freezing process, it would seem that the water-binding property of the colloidal system should have been materially decreased. Apparently the freezing process has in some manner produced an agglomeration of the colloidal particles, with no great effect upon the capillarity of the substance. The first freezing cycle (Sample 11)reduced the apparent “bound” water a t -6°C from 21.18% to 19.007C. A second freezing cycle reduced this only slightly and two additional freezing cycles were practically without effect insofar as reducing the amount of “bound” water is concerned I n the activated silica gel studies, no apparent disruption of the capillaries occurred, as is evidenced by the duplicability of the dilatometer readings upon repeated freezing and thawing. Foote and Saxton (13) report similar findings and note that this behavior of a silica gel system is not paralleled by the behavior of aluminum and ferric hydroxide gels. The General Behamor of Aqueous Collozdal Systems upon Exposure to Low Temperature.-It is widely accepted that with aqueous colloidal systems there is a definite temperature, below the freezing point of the system, a t which all of the “free-water’’ will be frozen. This conception is held especially in the field of biological investigation. With biological material this temperature has been arbitrarily selected as about - 20.0°C [ R ~ b n e rThoenes,60 ,~~ Robinson48,49,50],apparently on the assumption that all of the “free” water would be frozen a t this temperature and that the temperature was not sufficientlylow to seriously alter the free-bound water ratio.


433

Our experimental data indicate that all temperatures below that of the initial freezing of systems of gelatin-water are temperatures at which all of the “free-water” but none of the “bound-water” is frozen. With the inorganic hydrogels there was no temperature within the range of o.o°C to -5o.oOC a t which a lower temperature did not cause the crystallization of additional quantities of ice. Additional studies must be carried out before generalizations can be made as to the behavior of other aqueous colloidal systems upon exposure to low temperature. I t seems probable, however, that biological tissue and related systems would behave more nearly like the gelatin system than silica or ferric hydroxide gels. The results of this study further emphasize the importance of the length of the time interval during which the freezing of a colloidal system takes place. At temperatures decidedly lower than the freezing point of the system, the gel-ice equilibrium is rapidly established. At temperatures near the freezing point of the system the slow rate of ice formation is responsible for the slow establishment of phase equilibrium. Summwv Dilatometric studies have been made of the state of the water in certain aqueous colloidal systems as affected by exposure to temperatures ranging from o.o°C to - 50.0°C. The systems studied were: gelatin gels of different concentrations, the thick portion of egg white from fresh eggs, mixtures of different concentrations of activated silica gel and water, and hydrated ferric hydroxide. The low temperatures utilized in these studies were easily produced and maintained by proper mixtures of alcohol and solid carbon dioxide. The temperatures noted were measured by means of a standardized thermometer. 2. The colloidal systems studied fall into two general classes: (a) those in which the freezing process is completely reversible; and (b) those in which the freezing is an irreversible or only partially reversible process. 3. In the studies made upon gelatin it has been found that all the water which could be frozen within the temperature range of o.o°C to - 50.0°C was frozen a t the recorded initial freezing temperature (approx. - 6’C) if sufficient time was permitted for establishment of equilibrium. After equilibrium establishment at this point, exposure to much lower temperatures was without effect upon the quantity of water which remained in the unfrozen condition (“bound” water). 4. The rate of establishment of ice-gel equilibrium was found to be very slow, the rate being influenced greatly by (a) the concentration of the gel, and (b) the temperature of exposure. This fact emphasizes the importance of the consideration of time as a factor influencing the quantity of water which will be frozen a t any given temperature. 5 . Water-binding in the system, gelatin-water, appears to be an adsorp tion reaction. A logarithmic relationship was found to exist between the conI.

43 4

I V A N D. JOXES AND ROSS A I K E N GORTNER

centration of the gelatin gel and the grams of water which were “bound” per gram dry gelatin. In our studies 0.70 gram of water was “bound” per gram dry gelatin in a 32% gelatin gel and 4.675 grams of water per gram dry gelatin were “bound” in a zclc gelatin gel. 6. The freezing behavior of the thick portion of fresh egg white appears to be a completely reversible process, in this respect paralleling the gelatin systems. 7 . The form and the slope of the dilatometric curves indicated that no disruption of capillary spaces in silica gel occurred when silica gel-water systems were frozen. The freezing of colloidal ferric hydroxide resulted in a relatively small decrease in “capillary water.” 8. The quantity of “capillary-water” which could be frozen from ferric hydroxide and activated silica gel was proportional to the temperature a t which the sample was exposed. With ferric hydroxide and with ~ 5 7 activated ~ silica gel mixtures, this quantity of “capillary-water” frozen appeared to be directly proportional to the lowering of the temperature. With 60%) and 80.5% mixtures of activated silica gel this relationship held approximately to temperatures as low as -3oOC. At temperatures lower than -3o.o0C “capillary-water” froze at a more rapid rate than at temperatures above this point. 9. I t is postulated that the effect of freezing upon colloidal ferric hydroxide and similar substances is to cause in some manner an agglomeration or flocculation of the colloidal particles without materially reducing the capillarity exhibited by the substance. IO. The dilatometric method has proven a useful technic for “bound” water studies. The values for “bound” water which we have obtained in our work appear to approximate the values obtained by other workers using Rubner’s (Thoenes) calorimetric method on similar materials. 11. I t is emphasized that “bound” water is an indeterminate term, and that “bound” water values as experimentally determined may be expected to vary from system t o system, the variation being due to many factors, not the least of which is the method selected for measurement. If biological cells and tissues are similar in their behavior to gelatin and (probably) to the thick portion of egg white, then “bound” water is a measurable entity and (using dilatometric procedure) has a constant value a t least at temperatures between - 6’ and - jo°C.

Literature cited A. T. Barnes: Colloid Symposium Monograph, 3, 1 0 3 - 1 1 1 (1925). Colloidal Water and Ice. ‘ 0 . Bobertag, K. Feist and H. W. Fischer: Ber., 41, 3675-3679 (1908). t’ber das Ausfrieren von Hydrosolen. G. J. Bouyoucos: J. Agr. Res., 8, 195-217 (1917). Measurement of the Inactive or Gnfree Moisture in the Soil by Means of the Dilatometer Method. G. J. Bouyoucos: Mich. A r Expt. Sta. Tech. Bull. No. 36, (1917). Classification and Measurement of the Different kbrms of Water in the Soil by means of the Dilatometer Method. ‘D. R. Briggs: J. Phys. Chem., 36, 367 (1932).


43 5

G. Bruni: Ber., 42, 563-565 (1909). uber das Ausfrieren von Gallerten. ’ E. H. Callow: R o c . Roy. Soc., 10SA, 307-323 (1925). Ice Crystallization through Super-cooled Gelatin Gels. 8 J. W. Crist: Mich. Agr. Expt,. Sta. Tech. Bull. No. 74 (1926). Effect of Nutrient Conditions on Colloidal Properties of Certain Vegetoble Craps. 9 S. Dunn: Science, (2) 69, 359-360 (1929). Carbon Dioxide Ice as a Laboratory Ref rigerant . 1 0 H. A. Fells, and J. B. Firth: J. Phys. Chem., 31, 1230-1236(1927).Function of Water present in Silicic Acid Gel. Structure of Silicic Acid Gel. 11 H. W. Fischer: Kolloid-Z., 8, 291-302(1911). Daa Wasser im Plasma. 12 H. W. Fischer and 0. Bobertag: Biochem. Z., 18, 58-94 (1909). Uber das Ausfrieren von Gelen. 1s H. W. Foote and B. Saxton: J. Am. Chem. SOC.,38, 588-609 (1916). The Effect of Freezing on Certain Inorganic Hydrogels. 14 H. W. Foote and B. Saxton: J . Am. Chem. SOC.,39,627-630(1917).The Freezing of Water absorbed in Lamp Black. 16 H. W. Foote and B. Saxton: J. Am. Chem. SOC.,39, 1103-1125 (1917).The Effect of Freezing on Certain Inorganic Hydrosols. 16 H. Freundlich: “Colloid and Capillary Chemistry” (1926). 17R.A. Gortner: “Outlines of Biochemistry” (1929). 13 R. A. Gortner: Trans. Faraday Soc., 26, 678-686(note especially the a pendnd “Discussion” pp. 686-704) (1930). The State of Water in Colloidal and Living &stems. 19 W. B. Hardy: J. Phys. Chem., 4, 235-253 (1900). The Conditions which determine the Stability of Irreversible Hydrosols. 2 0 W. B. Hardy: J. Phye. Chem., 4, 253-273 (1900). On the Mechanism 01 Gelation in Reversible Colloidal Systems. 11 W. B. Hardy: Proc. Roy. SOC.,112A, 47-61 (1926). A Microscopic Study of the Freezing of a Gel. Unpuhlished work. Cited by Gortner (16)p. 23.5. 2f J. A. Harris, et. al. 23 A. V. Hill: Proc. Roy. Soc., 106B,477-505 (1930). The State of Water in Musc:e and Blood and the Osmotic Behavior of Muscle. *4 A. V. Hill and P. S. Kupalov: Proc. Roy. SOC., 106B, 445-477 (1930). The Vapor Pressure of Muscle. 26 International Critical Tables, 3, 27-29 (1928). 23 International Critical Tables, 3, 43 (1928). 2 7 H. R. Kruyt: “Colloids” (1929). ** H. R. Kruyt and K. C.,Winkler: 2. anorg. allgem. Chem., 188,200-204 (1930). c b e r den Einfluss hydratierter Kolloide auf die Gefrierpunktserniedrigung. 2 9 A. Kuhn: Kolloid-Z., 35, 275-294 (1924). uberblick unser jetzigen Kentnisse uber Wasaerbindung in Kolloiden. 30 R. E. Liesegang: Kolloid-Z., Flora, 96,523-524(1906).tfber daaErfrierender Pflanzen. 81 R. V. Lott: Mo. A r. Expt. Sta. Rsh. Bull. No. 95. Correlation of Chemical Composition with Hardiness in Brambles. 32 A. Lottermow: Ber., 41, 3976-3979 (1908). Uber das Ausfrieren van Hydrosolen. M. H. McCool and C. E. Millar: Bot. Gaz., 70, 317-319(1920). Use of the Dilatometer in studying Soil and Plant Relationships. 34 J. McGavack and W. A. Patrick: J.,Am. Chem. SOC.,42, 946-978 (1920). The Adsorption of Sulfur Dioxide by the Gel of Silicic Acid. Jb J. H. Martin: J. Agr. Res., 35, 493-535 (1927). Comparative Studies of Winter Hardiness ip Wheat. 38 H. Molisch: “Untersuchungen uber daa Erfrieren der Pflanzen” (I 897); Sitxungsber. Akad. Wiss. Wien, 105,Abt. I (1896).Das Erfrieren von Pflanzen bei Temperaturen uber dem Eispunkt. Cited by H. W. Fischer. (II). Originals not seen. T. Moran: Proc. Roy. SOC.,112A, 30-46 (1926). The Freezing of Gelatin Gel. 38 Muller-Thurgau: Landw. Jahrb., 9, 133-189(1880);15, 453-610 (1886). uber daa Gefrieren und Erfrieren der Pflanzen. 39R. Newton: Univ. Alberta Agr. Rsh. Bull., No. I (1923). The Nature and Practical Measurement of Frost Resistance in Winter Wheat. ‘0 R. Newton: J. Agr. Sci., 12, 1-19(1922). A Comparative Study of Winter Wheat Varieties with Especial Reference to Winter Killing. 8

43 6


4 1 R. Newton: J. Agr. Sci., 14, 178-191 (1924). Colloidal Properties of Winter Wheat Plants in Relation to Frost Resistance. 4 2 R. Newton and W. H. Cook: Canad. J. Res., 3, 560-578 (1930). The Bound Water of Wheat-Flour Suspensions. 4 3 R. Newton and R. A. Gortner: Bot. Gaz., 74, 442-446 (1922). A Method for estimating the Hydrophilic Colloid Content of Expressed Tissue Fluids. 4 4 R. Newton and W.McK. Martin: Can. J. Res., 3, 336-427 (1930). Physico-Chemical Studies on the Nature of Drought Resistance in Crop Plants. 4 b F. W. Parker: J. Am. Chem. SOC..43, 1011-1018 ( 1 9 2 1 ) . The Effect of Finely Divided Material on the Freezing Points of n’ater, Benzene and Xitro-Benzene. 4 8 W. A. Patrick and N. F. Ebcrman: J. Phys. Chem., 29, 220-228 (1925). Studies in Adsorption from Solutions from the Standpoint of Capillarit,y. 11. .?,W. A. Patrick and D. C. Jones: J. Phys. Chem., 29, 1 - 1 0 (1925). Studies in the Adsorption from Solution from the Standpoint of Capillarity. I. 4 8 W. Robinson: J . Econ. Entom., 20, 80-88 (1927). Water-Binding Capacity of Colloids, A Definite Factor in Winter Hardiness of Insects. 4 9 W. Robinson: Colloid Symposium Monograph, 5, 199-218 (1928). Relation of Hpdrophilic Colloids to Winter Hardiness of Insects. 5 0 W. Robinson: J. Econ. Entom., 21, 897-902 (1928). Water Conservation in Insects. 51 J.,T. Rosa, Jr.: Mo. Agr. Expt. Sta. Rsh. Bull., 48, ( 1 9 2 1 ) . Investigations on the Hardening Process in Vegetable Plants. 52 M. Rubner: Abh. preues. Akad. Wiss., Phys.-Math. Klasse, 1922, 3-70. tfber die Wasserhindung in Kolloiden mit besonderer Berucksichtigung des quergestreiften Muskels. 53 J. L. St. John: Unpublished Ms. Bound Water in Hydrophilic Colloids. s4 S.E. Shep,pard and S. Sweet: J. Ind. Eng. Chem., 13, 423-4’4 ( 1 9 2 1 ) . The Setting and Melting Points of Gelatins. 5s S.P. i.Sorensen: Compt. rend. trav. lab. Carlsberg, 12, 164-212 ( 1 9 1 7 ) . Studies on the Proteins. Special Report No. 7. London, H. M. Station56 W. Stiles: Food Investigation Board. ery Office. (1922). The Preservation of Food by Freezing, with Special Reference to Fish and Meat: A Study in General Physiology. T. Svedberg: J. Am. Chem. SOC.,46, 2673-2676 (1924). Density and Hydration in Gelatin Sols and Gels. 5 8 A. Taffel: J. Chem. SOC.,121, 1971-1984(1922). Thermal Expansion of Gelatin Gels. 69 G. Tammann: Z. physik. Chem., 23, 326-328 (1897). uber die Erstarrungsgeschwindigkeit. (cf. also J. Friedlander and G. Tammann: Z. physik. Chem., 24, 152-159 (1897)). Uber die Krystallisationsgeschwindigkeit; G. Tammann: “Kristallisieren und Schmelzen, 131-148. (1903). F. Thoenes: Biochem. Z., 157, 17 176 (1925). Untersuchungen zur Frageder Wasserbindung in Kolloiden und tierischen d k e b e n . 61 J. H. Walton and A. Brann: J. Am. Chem. SOC.,38, 317-330 ( 1 9 1 6 ) . The Effect of Dissolved Substances on the Velocit,yof Crystallization of Water. (* R. L. Vanzetti: Atti. Inst. Veneto, Sci., 75, 26: (1915); 76, 287 ( 1 9 1 6 ) . Cited by Fells and Firth: J. Phys. Chem., 31, 1231 11927). (Original not seen). 63 Vogel: Gilbert’s Ann. Physik., 46, 137 (1820). u b e r die Veranderung, welche einige Stoffe des organischen Reiches beim Gefrieren erfahren.

Division of Agricultural Biochemistry, University of Minnesota, St. Paul, Minnesota.

Free and Bound Water in Elastic and Non-Elastic Gels - The Journal of

Recommend Documents